Notatki do odcinka
VoLo wraps VLA policies as asynchronous tools called by a language model agent that monitors execution context and triggers recovery, achieving 3–5× higher success rates on the RoboVoLo benchmark and real Franka hardware compared to π0.5 alone. The approach demonstrates a scalable architecture for robust long-horizon robot control via LM-level orchestration.
Strona: share.transistor.fm

